CHOSUN

Reactive and Proactive Resource Allocation for LoRa-Enabled IoT Applications

Metadata Downloads
Author(s)
아라사하드 파하드
Issued Date
2022
Keyword
LoRa, LoRaWAN, Internet of Things, Adaptive Data Rate, Resource Allocation, Artificial Intelligence
Abstract
최근 LoRa는 다수의 단말에 대한 장거리 연결 지원, 낮은 배포 비용 및 적은 에너지 소비로 인해 사물인터넷 (IoT)을 위한 통신기술이 되었다. LoRaWAN의 ADR (adaptive data rate)은 스마트 그리드 및 계량과 같은 정적 IoT 애플리케이션에 권장되는 단말에 자원을 할당하기 위해 널리 채택된 전략이다. ADR은 개별 단말의 SF (spreading factor)와 TP (transmit power)를 관리한다. 그러나 고밀집 네트워크 및 가변 채널 조건에서는 ADR 성능이 크게 저하될 수 있다. 또한 단말의 이동성으로 인해 전송 환경이 자주 변경 되어 단말과 게이트웨이 간의 신호 강도에 영향을 준다. LoRaWAN의 ADR은 적절한 SF를 적용하여 안정적이고 에너지 효율적인 통신 상태로 수렴하는데 많은 시간이 소모되기 때문에 단말이 이동하는 경우에 부적절하다. 또한, LoRaWAN 네트워크 서버에 의해 새로 구성된 SF 와 TP 같은 통신 파라미터는 이동 단말과 게이트웨이 간의 효율적인 통신을 보장하지 않는다. 즉, 단말이 이동하는 경우 전파 환경이 급격히 변하여 새로 발행된 ADR 명령이 더 이상 유효하지 않아 단말은 적응적이지 못한 SF와 TP의 사용으로 인해 패킷 전송이 실패할 수 있다. 이 패킷 손실은 단말이 재전송 절차를 수행하도록 하여 패킷 간 간섭으로 인한 추가적인 손실을 일으킨다. 따라서 본 논문에서는 적응적이지 못한 SF로 인한 패킷 손실 문제를 reacitve, proactive, hybrid와 같은 다양한 자원 할당 접근 방식을 통해 해결한다. reacitve 접근 방식에는 채널 적응형 SF 할당(C-ASFA), 가우시안-ADR (G-ADR) 및 지수 이동 평균-ADR (EMA-ADR)이 포함된다. Proactive 방식으로는 M-ASFA (mobility-aware SF assignment), RA-ARM (retransmission-assisted resource management ADR), AI-ERA (artificial intelligence-empowered resource allocation)가 있다. 마지막으로는 hybrid ADR (HADR)를 제안한다. 제안한 방법의 주요 목표는 패킷 전송 성공률을 높이고 에너지 소비를 줄이는 것이다. 시뮬레이션 결과는 제안한 ADR 방식에 향상된 성능을 보여준다.|In recent times, LoRa has become a de-facto technology for the Internet of Things (IoT) owing to its long-range connectivity support for a large number of end devices, low deployment cost, and ultra-low energy consumption. The adaptive data rate (ADR) in LoRaWAN is a widely adopted strategy for resource assignment to end devices recommended for static IoT applications, such as smart grid and metering. ADR manages the spreading factor (SF) and transmit power (TP) of an individual end device (ED). However, the ADR performance can be reduced significantly under a highly dense network and variable channel conditions. Also, the mobility of EDs causes frequent alterations in the topology, which influences the signal strength between EDs and a gateway (GW). ADR of LoRaWAN is unsuitable and inadequate when the EDs are mobile because it requires hours to converge to a stable and energy-efficient communication state by adapting an appropriate SF. The new SF and TP configured by the NS do not guarantee efficient communication between mobile ED and GW. In such a case, the propagation environment may change radically when an ADR command reaches the mobile ED, and the newly assigned parameters may no longer be valid. Hence, a new packet from this ED with newly adopted parameters may be lost owing to the inappropriate use of both SF and TP. This packet loss compels EDs to retransmit the packet, resulting in further packet loss due to interference. Therefore, this dissertation resolves the packet loss issue caused due to inappropriate SF using reactive, proactive, and hybrid paradigms. The reactive paradigms comprise channel-adaptive SF allocation (C-ASFA), Gaussian-ADR (G-ADR), and Exponential moving average-ADR (EMA-ADR). The proactive paradigms include mobility-aware SF assignment (M-ASFA), retransmission-assisted resource management (RA-ARM), and artificial intelligence-empowered resource allocation (AI-ERA). Finally, a hybrid-ADR (HADR) is presented. Simulation results showed an improved packet success ratio, energy consumption, and convergence period compared to state-of-the-art ADR schemes.
Alternative Title
LoRa 기반 사물인터넷 응용을 위한 Reactive 및 Proactive 자원 할당
Alternative Author(s)
Arshad farhad
Affiliation
조선대학교 일반대학원
Department
일반대학원 정보통신공학과
Advisor
변재영
Awarded Date
2022-02
Table Of Contents
1 Introduction 1
1.1 Motivation 1
1.2 Problem Statement 2
1.3 Research Objectives 2
1.4 Contributions of Dissertation 3
1.4.1 Reactive Resource Allocation 3
1.4.2 Proactive Resource Allocation 4
1.4.3 Hybrid Resource Allocation 4
1.5 Organization of Dissertation 4

2 LoRaWAN Overview and Background Study 5
2.1 LoRa 5
2.1.1 LoRa Modulation 5
2.1.2 Spreading Factor 6
2.1.3 Time-On-Air 7
2.2 LoRaWAN 8
2.2.1 Types of End Devices in LoRaWAN 8
2.2.2 Confirmed Mode 9
2.2.3 Unconfirmed Mode 10
2.3 Adaptive Data Rates in LoRaWAN 10
2.3.1 ED-Managed ADR in Confirmed Mode 11
2.3.2 ED-Managed ADR in Unconfirmed Mode 11
2.3.3 NS-managed ADR 12
2.3.4 Blind Adaptive Data Rate 14
2.4 Background Studies 14
2.4.1 Interference-Based Approaches 15
2.4.2 Link- and System-Based Approaches 16
2.4.3 Mathematical Model-Based Approaches 17
2.4.4 Improvements in Typical ADR Approaches 18
2.4.5 Improvements in Convergence Period Approaches 18
2.4.6 Artificial Intelligence-Based Approaches 20
2.5 Summary 21

3 Network Model and Problem Formulation 22
3.1 Assumptions and Constraints 22
3.1.1 Class A End Devices 22
3.1.2 Frequency Region 22
3.1.3 Duty Cycle Constraints 22
3.2 Key Performance Indicators Utilized in Dissertation 23
3.2.1 Uplink Packet Outcomes 23
3.2.2 Convergence period 24
3.2.3 Energy consumption 24
3.3 Network Model 24
3.3.1 Channel Model 24
3.3.2 Channel Performance Model 27
3.4 Problem Formulation 29
3.4.1 Single mobile ED 29
3.4.2 Massive mobile EDs 30
3.4.3 Findings in BADR and ADR 32
3.5 Summary 35

4 Reactive Resource Allocation 36
4.1 Channel-Adaptive Spreading Factor Allocation 36
4.1.1 Increment SF 36
4.1.2 Decrement SF 37
4.2 Gaussian-Based Adaptive Data Rate 37
4.2.1 Scope of the Proposed G-ADR 37
4.2.2 Working Procedure of the Proposed G-ADR 37
4.3 Exponential Moving Average-Based Adaptive Data Rate 41
4.3.1 Scope of the Proposed EMA-ADR 41
4.3.2 Working Procedure of the Proposed EMA-ADR 42
4.4 Summary 42

5 Proactive Resource Allocation 44
5.1 The Mobility-Aware Spreading Factor Allocation 44
5.1.1 Initial SF Allocation with Traffic Heterogeneity 44
5.1.2 Mobility-Aware SF Assignment Scheme 45
5.2 Retransmission-Assisted Resource Management ADR 47
5.2.1 R-ARM at ED side 47
5.2.2 R-ARM at NS side 49
5.2.3 Integration of R-ARM in LoRaWAN 52
5.3 Design of the Proposed Artificial Intelligence-Empowered Re-source Allocation Framework 52
5.3.1 Scope of the Proposed AI-ERA Framework 52
5.3.2 AI-ERA Framework: Offline Mode 53
5.3.3 AI-ERA Framework: Online Mode 56
5.3.4 Computational Complexity of the Proposed Deep Learn-ing Model 57
5.4 Summary 58

6 Hybrid Resource Allocation 59
6.1 Hybrid Adaptive Data Rate 59
6.1.1 Computing d0 59
6.1.2 ADR Selection 59
6.2 Summary 61

7 Experimental Results 62
7.1 Application Scenario 62
7.2 Simulation Environment 63
7.3 Experimental Analysis of Reactive Resource Allocation 63
7.3.1 Analysis of C-ASFA 64
7.3.2 Analysis of G-ADR and EMA-ADR 64
7.4 Experimental Analysis of Proactive Resource Allocation 70
7.4.1 Analysis of M-ASFA 70
7.4.2 Analysis of R-ARM 74
7.4.3 Analysis of AI-ERA Framework 79
7.5 Experimental Analysis of Hybrid Resource Allocation 87
7.5.1 Analysis of HADR 87
7.6 Summary 92

8 Conclusions and Future Directions 93
8.1 Conclusions 93
8.2 Future directions 93

List of Publications 94

Biblography 96
Degree
Doctor
Publisher
조선대학교 대학원
Citation
아라사하드 파하드. (2022). Reactive and Proactive Resource Allocation for LoRa-Enabled IoT Applications.
Type
Dissertation
URI
https://oak.chosun.ac.kr/handle/2020.oak/17183
http://chosun.dcollection.net/common/orgView/200000606061
Appears in Collections:
General Graduate School > 4. Theses(Ph.D)
Authorize & License
  • AuthorizeOpen
  • Embargo2022-02-25
Files in This Item:

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.